For a data bank to be effective in clinical research, there must be a clear delineation of the research questions that will be addressed. An initial step, then, in proposing a data bank is developing a list of appropriate research questions and hypotheses. A broad set of research questions for a potential data bank in epilepsy has been suggested.